A bounds inference method for vector-based memoization

  • Authors:
  • Wei-Ngan Chin;Masami Hagiya

  • Affiliations:
  • National University of Singapore;University of Tokyo

  • Venue:
  • ICFP '97 Proceedings of the second ACM SIGPLAN international conference on Functional programming
  • Year:
  • 1997

Quantified Score

Hi-index 0.00

Visualization

Abstract

The dynamic-sized tabulation method can be used to eliminate redundant calls for certain classes of recursive programs. An innovative aspect of the method is the use of lambda abstractions that may subsequently be converted to bounded vectors, in order to share redundant calls via vector lookup.To facilitate this conversion to vector form, we propose a new inference method to conservatively determine the bounds for arithmetic parameters of recursive functions. Suitable techniques for inferring the safe bounds of these parameters are introduced, together with supporting transformations. The resulting method can obtain efficient vector-based programs without the need for run-time bounds checking.